Aleš Zamuda

This article presents a high-performance-computing differential-evolution-based hyperparameter optimization automated workflow (AutoDEHypO), which is deployed on a petascale supercomputer and utilizes multiple GPUs to execute a specialized fitness function for machine learning (ML). The workflow is designed for operational analytics of energy efficiency. In this differential evolution (DE) optimization use case, we analyze how energy efficiently the DE algorithm performs with different DE strategies and ML models. The workflow analysis considers key factors such as DE strategies and automated use case configurations, such as an ML model architecture and dataset, while monitoring both the achieved accuracy and the utilization of computing resources, such as the elapsed time and consumed energy. While the efficiency of a chosen DE strategy is assessed based on a multi-label supervised ML accuracy, operational data about the consumption of resources of individual completed jobs obtained from a Slurm database are reported. To demonstrate the impact on energy efficiency, using our analysis workflow, we visualize the obtained operational data and aggregate them with statistical tests that compare and group the energy efficiency of the DE strategies applied in the ML models.

This chapter describes a range of foundational concepts in differential evolution (DE) algorithm, including distance-based success history differential evolution (DISH), and then tackles some challenges from real-world applications (RWAs) of DE. The DISH algorithm is described in more detail on how it self-adapts control parameters and how it is applied on evolutionary computation challenges, including recent 100-digit challenges. The chapter then continues with some RWA applications of self-adaptive DE by listing outcomes from benchmarks prepared in robotics, computer animation, energy, and document understanding, as well as other benchmarks from competitions on evolutionary computation. In robotics, a success history DE for underwater glider path planning is defined, in computer animation, a mixed-integer multi-objective optimization DE for tree geometry animation is listed, in energy, parallel constrained DE optimization for hydro-thermal scheduling, and in document understanding, a discrete constrained DE for text summarization. Main methods and results demonstrating their applicability are provided.

This paper presents a deployment of DAPHNE (Integrated Data Analysis Pipelines for Large-Scale Data Management, HPC and Machine Learning) for Computational Intelligence (CI). As CI progresses, e.g. through effort in human societies like the IEEE Computational Intelligence Society (CIS), the latest advances have become more widely popularized. Likewise, with the accessibility of supercomputing resources to advance CI, like with infrastructures such as EuroHPC machines including Vega, this paper demonstrates deployment of Randomised Optimisation Algorithms (ROAs) as a subset of CI algorithms, on EuroHPC Vega. As the Vega is a cluster of distributed nodes, a successful computation on several nodes at once in parallel and concurrently, requires an orchestrated deployment across the nodes, batched using Slurm in this paper. An example ROA benchmarking scenario is then presented in this scope.

Postavitev in upravljanje računalniških oblakov: Zbirka vaj in nalog z rešitvami. Predmet Postavitev in upravljanje računalniških oblakov je vključen v bolonjski študijski program Računalništvo in informacijske tehnologije kot redna študijska enota. Ta dokument predstavlja študijsko gradivo za računalniške vaje iz tega predmeta. Po uvodu so na voljo posamezne naloge z rešitvami računalniških vaj v naslednjih poglavjih. Podan je tudi seznam arhiviranih datotek, ki vsebujejo računalniško obliko pripravljenih rešitev.

Underwater gliders are energy-efficient vehicles that rely on changes in buoyancy in order to convert up and down movement into forward displacement. These vehicles are conceived as multi-sensor platforms, and can be used to collect ocean data for long periods in wide range areas. This endurance is achieved at the cost of low speed, which requires extensive planning to ensure vehicle safety and mission success, particularly when dealing with strong ocean currents. As gliders are often involved on missions that pursue multiple objectives (track events, reach a target point, avoid obstacles, sample specified areas, save energy), path planning requires a way to deal with several constraints at the same time; this makes glider path planning a multi-objective (MO) optimization problem. In this work, we analyse the usage of the non-dominated sorting genetic algorithm II (NSGA-II) to tackle a MO glider path planning application on a complex environment integrating 3D and time varying ocean currents. Multiple experiments using a glider kinematic simulator coupled with NSGA-II, combining different control parameters were carried out, to find the best parameter configuration that provided suitable paths for the desired mission. Ultimately, the system described in this work was able to optimize multi-objective trajectories, providing non dominated solutions. Such a planning tool could be of great interest in real mission planning, to assist glider pilots in selecting the most convenient paths for the vehicle, taking into account ocean forecasts and particular characteristics of the deployment location.